القائمة
Voice sampling — study how Peter actually talks in conversations (not published pages). Captures rhythm, openings, vocabulary, humor. Accumulates observations over sessions.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | voice-sample |
| description | Voice sampling — study how Peter actually talks in conversations (not published pages). Captures rhythm, openings, vocabulary, humor. Accumulates observations over sessions. |
| user_invocable | true |
Before writing copy, you need to know how Peter actually sounds. Not how his published pages sound — how HE sounds when he's explaining something in conversation. This skill searches his conversation history, extracts voice patterns, and accumulates observations over time.
The voice sample is a tuning fork, not a style guide. After writing a draft, read it back and ask: does this sound like the person in those conversations? Not a polished version of him. Him.
/voice-sample [topic] # Sample how Peter talks about a specific topic
/voice-sample # General voice sampling (broader search)
memory/voice-sample.mdN/A — this skill observes, it doesn't evaluate against other skills.
N/A — this is an observation tool. It accumulates patterns, not verdicts.
Filter for Peter's words, not AI responses. Search in this order:
How Peter is talking to you right now. This is the freshest, most natural sample.
Location: /home/peter/.claude/projects/-home-peter-homelab-projects-active-petersalvato-com/*.jsonl
Filter: Look for "role":"human" entries.
grep -l "[topic]" /home/peter/.claude/projects/-home-peter-homelab-projects-active-petersalvato-com/*.jsonl
Then extract Peter's messages from matching sessions.
/pkb search "[topic]"
The PKB indexes 52,000+ documents across all sources. If available, use it for broad searches.
Location: /home/peter/homelab/knowledge/exports/JSON/conversations.json
Filter for Peter's messages (the human turns, not the assistant turns).
Location: /home/peter/homelab/knowledge/exports/consolidated_exports/consolidated_clean/
These are markdown — harder to filter speaker, but still useful for patterns.
From conversations, NOT from published pages:
Print observations in conversation. Also accumulate in the persistent memory file.
# Voice Sample — [topic or "General"]
## Rhythm
[How he strings sentences together]
## Opening Moves
[How he starts explaining — with examples quoted from conversations]
## Vocabulary
- Reaches for: [words he uses naturally]
- Never uses: [words absent from his speech]
## Tone
[Dry/warm/matter-of-fact/self-deprecating — with examples]
## Patterns
[Recurring structures, pet phrases, qualifiers]
## Example Quotes
[3-5 actual quotes from conversations that capture his voice]
Append observations to:
/home/peter/.claude/projects/-home-peter-homelab-projects-active-petersalvato-com/memory/voice-sample.md
Each session that does copy work should ADD to this file, not replace it. Over time it becomes a detailed fingerprint. Date-stamp each addition.
Format for persistent entries:
## [DATE] — [topic or context]
- [observation]
- [observation]
- Quote: "[actual quote]"
The Millman x Craftsman protocol (in voice/copywriting-protocol.md) defines the register: personal stakes through action, material vocabulary, no hype.
The voice sample defines the person inside that register. Two different writers could follow the same protocol and produce different copy. The voice sample ensures the copy sounds like Peter specifically.
/copy edit — Phase 1, Step 2 (while searching ideation history)/copy write — Step 3Context-aware story extraction. Loads the maker's full context — identity, expertise, voice, published work — then reads the narrative architecture to understand the story being told and what's missing. Interviews the maker to fill gaps with real moments. Modular: works with any project.
Evaluate the IP constellation — dispatches baseline checks, then 9 lenses in parallel (Millman, Bierut, Appleton, Shaw, Peers, Victore, The Walk, The Walls). Coordinator skill.
Cross-page coherence. Do the three tiers form a coherent argument when read as a sequence? Read-only diagnostic.
Baseline checks — mechanical health of the constellation. Links, images, cross-refs, attribution, voice violations, SEO, tooltips. Read-only diagnostic.
Where lenses agree and disagree. Maps consensus vs. contradiction across audit results. Read-only diagnostic.
Copy verification checklist — 13-item pass/fail check on draft or published copy. Voice rules, grounding, Shaw check, steward voice. Read-only diagnostic.